The algorithm description mentions scalars $\alpha, \beta, \kappa$ which do not appear in the algorithm.
Example 19.4 sounds like sqrt(::Matrix) cannot be used to compute a matrix square root with A=BB'. However, in alg 19.5 sqrt is used instead of the earlier mentioned Cholesky decomposition. I assume that sqrt(A::Matrix) returns a symmetric matrix if A is symmetric positive definite in which case it doesn't matter, but maybe example 19.4 should reflect this.
The algorithm description mentions scalars
$\alpha, \beta, \kappa$
which do not appear in the algorithm.Example 19.4 sounds like
sqrt(::Matrix)
cannot be used to compute a matrix square root withA=BB'
. However, in alg 19.5sqrt
is used instead of the earlier mentioned Cholesky decomposition. I assume thatsqrt(A::Matrix)
returns a symmetric matrix ifA
is symmetric positive definite in which case it doesn't matter, but maybe example 19.4 should reflect this.